Two different approaches to automated mark up of emotions in text

Abstract

This paper presents two different approaches to automated marking up of texts with emotional labels. For the first approach a corpus of example texts previously annotated by human evaluators is mined for an initial assignment of emotional features to words. This results in a List of Emotional Words (LEW) which becomes a useful resource for later automated mark up. The mark up algorithm in this first approach mirrors closely the steps taken during feature extraction, employing for the actual assignment of emotional features a combination of the LEW resource and WordNet for knowledge-based expansion of words not occurring in LEW. The algorithm for automated mark up is tested against new text samples to test its coverage. The second approach mark up texts during their generation. We have a knowledge base which contains the necessary information for marking up the text. This information is related to actions and characters. The algorithm in this case employ the information of the knowledge database and decides the correct emotion for every sentence. The algorithm for automated mark up is tested against four different texts. The results of the two approaches are compared and discussed with respect to three main issues: relative adequacy of each one of the representations used, correctness and coverage of the proposed algorithms, and additional techniques and solutions that may be employed to improve the results. © 2007 Springer-Verlag London.